Conjunctival epithelial cells resist productive SARS-CoV-2 infection

Abstract

Conjunctival epithelial cells, which express viral-entry receptors angiotensin-converting enzyme 2 (ACE2) and transmembrane protease serine type 2 (TMPRSS2), constitute the largest exposed epithelium of the ocular surface tissue and may represent a relevant viral-entry route. To address this question, we generated an organotypic air-liquid-interface model of conjunctival epithelium, composed of basal, suprabasal, and superficial epithelial cells, and fibroblasts, which could be maintained successfully up to day 75 of differentiation. Using single-cell RNA sequencing (RNA-seq), with complementary imaging and virological assays, we observed that while all conjunctival cell types were permissive to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genome expression, a productive infection did not ensue. The early innate immune response to SARS-CoV-2 infection in conjunctival cells was characterised by a robust autocrine and paracrine NF-κB activity, without activation of antiviral interferon signalling. Collectively, these data enrich our understanding of SARS-CoV-2 infection at the human ocular surface, with potential implications for the design of preventive strategies and conjunctival transplantation.

Keywords: ACE2; IFN; NFKB; SARS-CoV-2; TMPRSS2; conjunctiva; conjunctival epithelium; ocular surface; productive infection.

Graphical abstract

Figure 1

Generation and characterization of the ALI conjunctival organotypic culture model (A) Schematic summary showing the key steps involved in generation of the ALI conjunctival organotypic culture model. (B) RNA expression of conjunctival epithelial cell markers in the human adult cornea and conjunctiva single-cell RNA-seq data reported recently by Collin et al. (2021a). Raw expression values were normalized, log transformed, and summarized. The size of the dots indicates the proportion of cells, while the color indicates the mean expression. (C) Quantitative RT-PCR analysis showing expression of conjunctival cell-type-specific markers and SARS-CoV-2 entry factors (ACE2 and TMPRSS2). Data shown as mean ± SEM, n = 3–14 experimental repeats in three different donors, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, one way ANOVA with Tukey’s multiple comparisons. Conj, conjunctiva. (D) Whole-mount immunofluorescence analysis showing expression of conjunctival epithelial marker (KRT13), superficial conjunctival epithelial marker (KRT4, KRT7), goblet cell marker (MUC5AC), mucin-producing cells (MUC4), and SARS-CoV-2 entry factors (ACE2, TMPRSS2) in day 15 ALI conjunctival organotypic culture models (representative of repeat experiments in three different donor conjunctival ALI cultures). Hoe, Hoescht. Scale bars 50 μm.

Figure 2

Characterization of ALI conjunctival organotypic culture model at day 30 of differentiation by immunofluorescence and single-cell RNA-seq (A) Immunofluorescence analysis showing co-expression of ACE2 and TMPRSS2 in the superficial layer of the ALI conjunctival organotypic model. KRT7 and KRT4 were predominantly located in the superficial layer, while MUC4 was detected throughout (representative of repeat experiments in three different donor conjunctival ALI cultures). Hoe, Hoescht. Scale bars 50 μm. (B) Presence of neutral mucins (magenta) in whole-mount stains of ALI conjunctival organotypic model (representative of repeat experiments in three different donor conjunctival ALI cultures). Scale bars 50 μm. (C) Uniform manifold approximation and projection (UMAP) visualization of scRNA-seq data from conjunctival ALI cultures (8,202 cells from three different donors) showing the presence of basal (CjB), suprabasal (CjSB), and superficial conjunctival (CjS) epithelium and fibroblasts (Fibs). Expression of key epithelial markers and SARS-CoV-2 entry factors, ACE2 and TMPRSS2, are shown as superimposed single gene-expression plots on the UMAP.

Figure 3

SARS-CoV-2 infection of day 30 human ALI conjunctival organotypic culture (A and B) Quantitative RT-PCR expression of nucleocapsid (N) gene (normalized to the housekeeper RNASEP) and subgenomic N RNA (normalized to GAPDH) from 0–72 hpi. Data shown as mean ± SEM, n = 3–7 experimental repeats, 3 different donors, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, one way ANOVA with Dunnett’s multiple comparisons to 0 hpi. (C) Representative western blot showing the expression of SARS-CoV-2 spike (S; shown by asterisk) and cleaved S2 protein expression (shown by two asterisks) in the nasal and conjunctival ALI organotypic culture models. GAPDH was used as loading control (representative of repeat experiment in 3 donors). (D) Release of infectious viral particles was determined by plaque assay using apical washings from 2–72 hpi. Data shown as mean ± SEM, n = 2–5, 3 different donors, ^∗p < 0.05, ^∗∗p < 0.01, one way ANOVA with Dunnett’s multiple comparisons to 2 hpi. (E) Immunofluorescence analysis showing the presence of infected cells marked by ACE2 and S co-expression. A few mucin-secreting cells are also infected by SARS-CoV-2, as shown by co-expression of MUC4 and MUC5AC with S (white arrows). A panel of mock-infected cells is shown on the right-hand side panel, and white arrows indicate MUC5AC-secreting cells (representative of repeat experiment in 3 donors). Hoe, Hoescht. Scale bars 50 μm.

Figure 4

Single-cell RNA-seq analyses at 24 hpi reveal broad but low tropism of SARS-CoV-2 in the ALI conjunctival organotypic culture model (A) UMAP visualization of scRNA-seq data from mock and infected conjunctival ALI cultures (15,821 cells from three different donors, mock and SARS-CoV-2 infected) showing the presence of CjB, CjSB, and CjS epithelium and Fibs. A smaller UMAP on the right bottom corner shows the overlap between SARS-CoV-2-exposed and unexposed cultures. (B and C) Expression of S and N transcripts shown as superimposed single gene-expression plots on the UMAP. (D) Relative proportion of infected cell types (CjB, CjSB, CjS, Fibs) based on expression of any viral transcript. (E) Dot plot demonstrating expression of key chemokine marker upregulated in response to SARS-CoV-2 infection in all cell types, with intensity demonstrated by color and size of the dot representing the proportion of cells expressing the marker.

Figure 5

SARS-CoV-2 infection in conjunctival cells is characterized by robust autocrine and paracrine NF-κB activity (A and B) Representative network analysis of predicted regulators in the SARS-CoV-2-infected cells in the CjB (A) and CjS (B) epithelial cells. Differentially expressed genes between infected and mock cells within the CjB and CjS epithelium cluster were generated using the Seurat FindMarkers function. IPA upstream regulator analysis was used to predict upstream transcriptional regulators from this gene list, using the Ingenuity Knowledge Base to create mechanistic networks. (C and D) NF-κB target (C) and IFN-stimulated gene (ISG) expression (D) in infected, bystander, and mock in the CjB, CjSB, and CjS epithelium and fibroblasts. Gene set scores greater than zero suggest expression levels higher than background gene expression. The bottom and the top of the boxes correspond to the 25th (Q1) and 75th (Q3) percentiles, and the internal band is the 50th percentile (median). The plot whisker minimum is calculated as Q1 -1.5 × interquartile range (IQR) and the maximum as Q3 +1.5 × IQR. IQR, interquartile range. Outside points correspond to potential outliers.

Figure 6

Evasion of IFN response in infected conjunctival epithelial cells by SARS-CoV-2 (A–C) Representative network analysis of predicted regulators in the bystander cells in the CjB (A), superficial (B), and CjSB (C) epithelial cells. Differentially expressed genes between bystander and mock-infected cells within the CjB, CjSB, and CjS epithelial cells were generated using the Seurat FindMarkers function. IPA upstream regulator analysis was used to predict upstream transcriptional regulators from this gene list, using the Ingenuity Knowledge Base to create mechanistic networks.

Figure 7

Transcriptomic comparison of conjunctival and nasal epithelium response to SARS-CoV-2 infection (A and B) UMAPs showing the integration of single-cell RNA-seq data of CjS and nasal secretory and ciliated epithelial cells exposed and unexposed to SARS-CoV-2. (C) Differentially expressed genes between unexposed (mock) CjS and nasal secretory and ciliated epithelial cells were generated using the Seurat FindMarkers function. IPA upstream regulator analysis was used to predict upstream transcriptional regulators from this gene list, using the Ingenuity Knowledge Base to create mechanistic networks. (D) An IPA comparison analysis of canonical pathways identified by a comparison between CjS and secretory and ciliated nasal epithelial cells within mock (M), bystander (B), and infected (I) groups. The results are filtered to show any significant pathways with an absolute Z score of 2 or more. Orange represents pathways with a positive Z score that are activated, and blue indicates negative Z scores that are deactivated in CjS compared with secretory and ciliated nasal epithelial cells. A dot indicates pathways with an absolute Z score of <2.